| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 1.67 MB | Adobe PDF |
Authors
Abstract(s)
Nos últimos anos, têm sido desenvolvidos vários modelos de escolha
discreta para tentar descrever o comportamento dos consumidores em
mercados de produtos diferenciados. Estes modelos são hoje utilizados por
várias empresas para fazerem previsões da procura, e com isso conseguirem
definir estratégias que lhes permitam ganhar vantagem competitiva.
Neste trabalho, aplicou-se um tipo de modelo de escolha discreta – o modelo
logit – para estimar a procura de um conjunto de produtos diferenciados,
nomeadamente desparasitantes externos de cão, após implementação de uma
campanha de marketing, nas várias unidades pertencentes à empresa
Onevetgroup.
Para poder efetuar a estimativa pretendida pelo modelo logit, foram
consideradas as seguintes variáveis: as quantidades e os preços de
desparasitantes externos vendidos entre Dezembro 2013 e Outubro 2015, as
características dos produtos, e o mercado potencial. De seguida, fazendo uma
regressão através do método dos mínimos quadrados de lnsj – lns0 em função
dos preços e características dos produtos, foi possível determinar os coeficientes
necessários para prever a quantidade de desparasitantes externos vendidos em
CV Algés em Outubro 2015.
Nesse período, os resultados demonstraram que as previsões de venda de
desparasitantes foram próximas da realidade. Como a qualidade da previsão foi
satisfatória, foi então possível medir o impacto que uma campanha de
marketing poderia ter nas vendas da empresa. Concluiu-se que esta campanha
levaria a uma aumento das vendas, gerando consequentemente um aumento
das receitas (em cerca de 28%). No entanto, é necessário replicar esta
metodologia nas restantes unidades pertencentes à Onevetgroup, de modo a afirmar com maior grau de certeza que este modelo é apropriado para fazer
previsões na empresa.
In recent years several discrete choice models have been developed to try to describe consumer’s behavior in markets of differentiated product. These models are nowadays used by many companies to estimate demand, and thus to define strategies that enable them to gain competitive advantage. In the context of this dissertation, one type of discrete choice model was applied – the logit model – to estimate demand of differentiated products, namely external deworming for dogs, after the implementation of a marketing campaign, in the different units of the Onevetgroup company. The data required to consistently estimate the logit model consists of the following variables: quantities and prices of external deworming sold between December 2013 and October 2015, products characteristics, and the potential market. Then, using least squares regression of lnsj – lns0 in function of prices and products characteristics, it was possible to determine the coefficients to estimate the quantity of external deworming sold in CV Algés in October 2015. During this period, the results demonstrated that the deworming sales estimations were close to reality. The quality of the estimations was satisfying, which allowed measuring the impact of a marketing campaign on the company's sales. We concluded that this campaign would lead to an increase in sales, thereby generating increases in revenue (about 28 %). However, we need to replicate this methodology in the other units of Onevetgroup to affirm, with a higher degree of certainty, that this is an appropriate forecast model for the company.
In recent years several discrete choice models have been developed to try to describe consumer’s behavior in markets of differentiated product. These models are nowadays used by many companies to estimate demand, and thus to define strategies that enable them to gain competitive advantage. In the context of this dissertation, one type of discrete choice model was applied – the logit model – to estimate demand of differentiated products, namely external deworming for dogs, after the implementation of a marketing campaign, in the different units of the Onevetgroup company. The data required to consistently estimate the logit model consists of the following variables: quantities and prices of external deworming sold between December 2013 and October 2015, products characteristics, and the potential market. Then, using least squares regression of lnsj – lns0 in function of prices and products characteristics, it was possible to determine the coefficients to estimate the quantity of external deworming sold in CV Algés in October 2015. During this period, the results demonstrated that the deworming sales estimations were close to reality. The quality of the estimations was satisfying, which allowed measuring the impact of a marketing campaign on the company's sales. We concluded that this campaign would lead to an increase in sales, thereby generating increases in revenue (about 28 %). However, we need to replicate this methodology in the other units of Onevetgroup to affirm, with a higher degree of certainty, that this is an appropriate forecast model for the company.
Description
Keywords
Procura Produtos diferenciados Modelos de escolha discreta Modelo logit Previsão Demand Differentiated products Discrete choice models Logit model Estimation